Optimization of codes for the joint Viterbi detector decoder (JVDD)

The joint Viterbi detector decoder (JVDD) was recently proposed as an alternative to iterative detector/decoder systems (also known as turbo equalizers) used in communication applications today. Despite being conditionally optimal over coded additive white Gaussian noise (AWGN)/inter-symbol interference (ISI) channels, the JVDD suffers from an exponential growth in the number of survivors when used with existing codes such as random low-density parity check (LDPC) codes. This paper proposes a class of JVDD codes designed for the JVDD which mitigate this undesirable trait. This class of codes consists of 4 code designs — 1) fixed row weight, 2) fixed column weight, 3) proportionate row weight and 4) proportionate column weight. While the proposed JVDD codes do effectively bring about a reduction in complexity and enhance the performance of the JVDD, the design parameters of these codes require optimization. In this paper, a design of experiment (DOE) technique that employs the response surface methodology (RSM) is used as a tool to optimize the parameters of each of the proposed codes.